The coordination of dynamic neural activity within and between neural networks is believed to underlie normal cognitive processes. Conversely, cognitive deficits that occur following neurological insults may result from network discoordination. We hypothesized that cognitive outcome following febrile status epilepticus (FSE) depends on network efficacy within and between fields CA1 and CA3 to dynamically organize cell activity by theta phase. Control and FSE rats were trained to forage or perform an active avoidance spatial task. FSE rats were sorted by those that were able to reach task criterion (FSE-L) and those that could not (FSE-NL). FSE-NL CA1 place cells did not exhibit phase preference in either context and exhibited poor cross-theta interaction between CA1 and CA3. FSE-L and control CA1 place cells exhibited phase preference at peak theta that shifted during active avoidance to the same static phase preference observed in CA3. Temporal coordination of neuronal activity by theta phase may therefore explain variability in cognitive outcome following neurological insults in early development.

Evidence from animal models and patient data indicates that febrile status epilepticus (FSE) in early development can result in permanently diminished cognitive abilities. To understand the variability in cognitive outcome following FSE, we used MRI to measure dynamic brain metabolic responses to the induction of FSE in juvenile rats. We then compared these measurements to the ability to learn an active avoidance spatial task weeks later. T2 relaxation times were significantly lower in FSE rats that were task learners in comparison to FSE non-learners. While T2 time in whole brain held the greatest predictive power, T2 in hippocampus and basolateral amygdala were also excellent predictors. These signal differences in response to FSE indicate that rats that fail to meet metabolic and oxygen demand are more likely to develop spatial cognition deficits. Place cells from FSE non-learners had significantly larger firing fields and higher in-field firing rate than FSE learners and control animals and imply increased excitability in the pyramidal cells of FSE non-learners. These findings suggest a mechanistic cause for the spatial memory deficits in active avoidance and are relevant to other acute neurological insults in early development where cognitive outcome is a concern.

In working memory tasks, stimulus presentation induces a resetting of intracranial temporal lobe oscillations in multiple frequency bands. To further understand the functional relevance of this phenomenon, we investigated whether working memory performance depends on the phase precision of ongoing oscillations in the hippocampus. We recorded intra-hippocampal local field potentials in individuals performing a working memory task. Two types of trials were administered. For high memory trials presentation of a list of four letters ("List") was followed by a single letter memory probe ("Test"). Low memory load trials, consisting of four identical letters (AAAA) followed by a probe with the same letter (A), were interspersed. Significant phase locking of ongoing oscillations across trials, estimated by the Pairwise Phase Consistency Index (PPCI) was observed in delta (0.5-4 Hz), theta (5-7 Hz), and alpha (8-12 Hz) bands during stimulus presentation and recall but was increased in low memory load trials. Across patients however, higher delta PPCIs during recall in the left hippocampus were associated with faster reaction times. Because phase locking could also be interpreted as a consequence of a stimulus evoked potential, we performed event related potential analysis (ERP) and examined the relationship of ERP components with performance. We found that both amplitude and latency of late ERP components correlated with both reaction time and accuracy. We propose that, in the Sternberg task, phase locking of oscillations, or alternatively its ERP correlate, synchronizes networks within the hippocampus and connected structures that are involved in working memory.